Substitution Boundaries of AI Agents in Team Collaboration: An Empirical Analysis Based on Front-Line Employees' Perceptions
Year: 2026
Type: article
Study design: Observational Study
Abstract: With AI agents increasingly involved in teamwork, whether AI can substitute human collaboration has become a heated topic in academia. Large-model-driven agents are now capable of participating in team brainstorming, task coordination, and collaborative problem-solving, and knowledge-intensive industries are beginning to form a large number of human-machine hybrid teams. Existing research mostly relies on theoretical reasoning and lacks empirical perceptual data from front-line practitioners. Drawing on a questionnaire survey of 158 employees with cross-industry teamwork experience, this paper explores practitioners' perceptions regarding the substitution of human collaborative behaviors by AI. The results show that AI can only partially substitute procedural and rule-oriented collaborative tasks, whereas social-emotional communication, conflict mediation and value judgment remain irreplaceable human strengths. Position, work experience and AI familiarity jointly shape individual perceptions of the substitution boundary. Most practitioners regard AI as an auxiliary tool rather than a leader of team collaboration. This study provides a practical reference for enterprises applying AI collaboration tools.
Author: Ziqian Chang
Institution: Nanjing Foreign Language School
Language: English
Subfield: Social Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: hybrid
OpenAlex ID: W7225716263
